
log using "C:\Users\endresk\Dropbox\SSRI_KE\Mode Experiments\PSRM_Analysis.log", replace
use "C:\Users\endresk\Dropbox\SSRI_KE\Mode Experiments\PSRM_Final.dta"

// ISSUES MENTIONED

** online (Text page 13)
mean mip_online_problemcount
** video (Text page 13)
mean mip_ibrc_problemcount if mode==1
** in-person (Text page 13)
mean mip_ibrc_problemcount if mode==2

// FN 14 (page 13): percent volunteering more than one issue
** online
tab mip_online_morethanone
** video
tab mip_ibrc_morethanone if mode==1
** in-person
tab mip_ibrc_morethanone if mode==2

// Figure 2: Mode Differences in Mean Number of Issues Mentioned (page 14 and 15)

** Between Subject (Video v. In-person)
ttest mip_ibrc_problemcount, by(mode)
** Within Subject (Online v. Video)
ttest mip_ibrc_problemcount = mip_online_problemcount if mode==1
** Within Subject (Online v. In-person)
ttest mip_ibrc_problemcount = mip_online_problemcount if mode==2
** Within Subject (Online v. Combined)
ttest mip_ibrc_problemcount = mip_online_problemcount


** Issue Differences(Text page 14)
tab mip_diff



// ITEM NON RESPONSE


// Figure 3: Mode Differences in Percentage Flagged for Item Nonresponse (page 15 and 16)

** Between Subject (Video v. In-person)
ttest itemnonresponse_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest itemnonresponse_ibrc== itemnonresponse_online if mode==1
** Within Subject (Online v. In-person)
ttest itemnonresponse_ibrc== itemnonresponse_online if mode==2
** Within Subject (Online v. Combined)
ttest itemnonresponse_ibrc== itemnonresponse_online



// STRAIGHTLINING


// Figure 4: Mode Differences in Percentage Flagged for Straighlining (page 16 and 17)

** Between Subject (Video v. In-person)
ttest sl_any_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest sl_any_ibrc== sl_any_online if mode==1
** Within Subject (Online v. In-person)
ttest sl_any_ibrc== sl_any_online if mode==2
** Within Subject (Online v. Combined)
ttest sl_any_ibrc== sl_any_online

// FN 17 (page 16): American Identity Battery
** Within Subject (Online v. Combined)
ttest ai_ibrc_sl_flag== ai_online_sl_flag
** Between Subject (Video v. In-person)
ttest ai_ibrc_sl_flag, by(mode)
** Within Subject (Online v. Video)
ttest ai_ibrc_sl_flag== ai_online_sl_flag if mode==1
** Within Subject (Online v. In-person)
ttest ai_ibrc_sl_flag== ai_online_sl_flag if mode==2



// SOCIAL DESIRABILITY BIAS

// Figure 5: Modes Differences in Social Desirability Effects



// Racial Resentment (page 18 and 19 + Figure 5 on page 20)

** Between Subject (Video v. In-person)
ttest resentment_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest resentment_ibrc = resentment_online if mode==1
** Within Subject (Online v. In-person)
ttest resentment_ibrc = resentment_online if mode==2
** Within Subject (Online v. Combined)
ttest resentment_ibrc = resentment_online

** Percentage Moving in socially desirable direction (Text page 19)
tab resentment_diff




// Immigration (page 19 and 20 + Figure 5 on page 20)
** Between Subject (Video v. In-person)
ttest imm_battery_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest imm_battery_ibrc = imm_battery_online if mode==1
** Within Subject (Online v. In-person)
ttest imm_battery_ibrc = imm_battery_online if mode==2
** Within Subject (Online v. Combined)
ttest imm_battery_ibrc= imm_battery_online

** Percentage Moving in socially desirable direction (Text page 19 and 20)
tab imm_diff

// FN 19 (page 12) non-attitudinal items
tab income_consistency
tab news_consistency

// Figure 6: Modes Differences in Feeling Thermometer Scores (page 22)

// Blacks
** Between Subject (Video v. In-person)
ttest ft_blacks_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest ft_blacks_ibrc = ft_blacks_online if mode==1
** Within Subject (Online v. In-person)
ttest ft_blacks_ibrc = ft_blacks_online if mode==2
** Within Subject (Online v. Combined)
ttest ft_blacks_ibrc = ft_blacks_online

// Democrats
** Between Subject (Video v. In-person)
ttest ft_dems_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest ft_dems_ibrc=ft_dems_online if mode==1
** Within Subject (Online v. In-person)
ttest ft_dems_ibrc=ft_dems_online if mode==2
** Within Subject (Online v. Combined)
ttest ft_dems_ibrc=ft_dems_online

// Evangelicals
** Between Subject (Video v. In-person)
ttest ft_evanelicals_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest ft_evanelicals_ibrc = ft_evanelicals_online if mode==1
** Within Subject (Online v. In-person)
ttest ft_evanelicals_ibrc = ft_evanelicals_online if mode==2
** Within Subject (Online v. Combined)
ttest ft_evanelicals_ibrc = ft_evanelicals_online

// Gay men and lesbians
** Between Subject (Video v. In-person)
ttest ft_gays_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest ft_gays_ibrc = ft_gays_online if mode==1
** Within Subject (Online v. In-person)
ttest ft_gays_ibrc = ft_gays_online if mode==2
** Within Subject (Online v. Combined)
ttest ft_gays_ibrc = ft_gays_online

// Muslims
** Between Subject (Video v. In-person)
ttest ft_muslims_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest ft_muslims_ibrc = ft_muslims_online if mode==1
** Within Subject (Online v. In-person)
ttest ft_muslims_ibrc = ft_muslims_online if mode==2
** Within Subject (Online v. Combined)
ttest ft_muslims_ibrc = ft_muslims_online

// Republicans
** Between Subject (Video v. In-person)
ttest ft_reps_ibrc, by(mode)
** Within Subject (Online v. Video)
ttest ft_reps_ibrc = ft_reps_online if mode==1
** Within Subject (Online v. In-person)
ttest ft_reps_ibrc = ft_reps_online if mode==2
** Within Subject (Online v. Combined)
ttest ft_reps_ibrc = ft_reps_online

** FN 21 percent moving in socially desirable direction (page 21)
tab socdes_ftdems
tab socdes_ftreps
tab socdes_ftevan
tab socdes_ftmuslims
tab socdes_ftblacks
tab socdes_ftgays

// Participant satisfaction  (page 23)
ttest int_satis_paper_rev, by(mode)

// Percentage distracted during interviewer administered (page 23)
ttest int_distracted_paper, by(mode)

// Figure 7: Mean Evaluation of Interview Experience (page 23 and 24)
ttest paper_honestly, by(mode)
ttest paper_boring, by(mode)
ttest paper_interesting, by(mode)
ttest paper_long, by(mode)
ttest paper_personal, by(mode)
ttest paper_matter, by(mode)

// Interviewer evaluations (page 24)
ttest observ_distract_vid, by(mode)
ttest observe_pk_ibrc, by(mode)
ttest observe_honest_ibrc, by(mode)



// Table A1: Balance between interviewer-administered conditions
ttest interviewer1, by(mode)
ttest interviewer2, by(mode)
ttest interviewer3, by(mode)
ttest interviewer4, by(mode)

ttest age, by(mode)
ttest age_18_25, by(mode)
ttest age_26_39, by(mode)
ttest age_40, by(mode)

ttest registered_online, by(mode)
ttest democrat_online, by(mode)
ttest republican_online, by(mode)
ttest female, by(mode)
ttest educ_4year, by(mode)
ttest educ_gradprof, by(mode)
ttest race_white, by(mode)
ttest race_black, by(mode)
ttest race_asian, by(mode)


// Table A2: Data Quality Differences between modes
** Mean Issues mentioned
** Within Subject (Online v. Video)
ttest mip_ibrc_problemcount = mip_online_problemcount if mode==1
** Within Subject (Online v. In-person)
ttest mip_ibrc_problemcount = mip_online_problemcount if mode==2
** Within Subject (Online v. Combined)
ttest mip_ibrc_problemcount = mip_online_problemcount
** Between Subject (Video v. In-person)
ttest mip_ibrc_problemcount, by(mode)

** Percent flagged for Item Nonresponse
** Within Subject (Online v. Video)
ttest itemnonresponse_ibrc== itemnonresponse_online if mode==1
** Within Subject (Online v. In-person)
ttest itemnonresponse_ibrc== itemnonresponse_online if mode==2
** Within Subject (Online v. Combined)
ttest itemnonresponse_ibrc== itemnonresponse_online
** Between Subject (Video v. In-person)
ttest itemnonresponse_ibrc, by(mode)

** Percent flagged for "straightlining"
** Within Subject (Online v. Video)
ttest sl_any_ibrc== sl_any_online if mode==1
** Within Subject (Online v. In-person)
ttest sl_any_ibrc== sl_any_online if mode==2
** Within Subject (Online v. Combined)
ttest sl_any_ibrc== sl_any_online
** Between Subject (Video v. In-person)
ttest sl_any_ibrc, by(mode)

// Table A3: Multivariate Regression models for Data Quality Metrics
poisson mip_ibrc_problemcount i.video mip_online_problemcount i.male educ age  i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4
regress itemnonresponse_ibrc i.video i.itemnonresponse_online i.male educ age  i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4
regress sl_any_ibrc i.video i.sl_any_online i.male educ age i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4

// Table A4: Multivariate Regression models for Social Desirability
regress resentment_ibrc i.video resentment_online i.male educ age i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4 
regress imm_battery_ibrc i.video imm_battery_online i.male educ age i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4 

// Table A5: Multivariate Regression models for Feeling Thermometers
regress ft_blacks_ibrc i.video ft_blacks_online i.male educ age i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4
regress ft_evanelicals_ibrc i.video ft_evanelicals_online i.male educ age i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4
regress ft_gays_ibrc i.video ft_gays_online i.male educ age i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4
regress ft_muslims_ibrc i.video ft_muslims_online i.male educ age i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4
regress ft_dems_ibrc i.video ft_dems_online i.male educ age i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4
regress ft_reps_ibrc i.video ft_reps_online i.male educ age i.race_black i.race_asian i.hispanic i.interviewer2 i.interviewer3 i.interviewer4



// Table A6: Results Segmented by Time Duration Between Waves

** Racial resentment
** Below median
ttest resentment_ibrc = resentment_online if timegap_minutes<1559
** at or above median
ttest resentment_ibrc = resentment_online if timegap_minutes>=1559

** Immigrant
** Below median
ttest imm_battery_ibrc= imm_battery_online if timegap_minutes<1559
** at or above median
ttest imm_battery_ibrc= imm_battery_online if timegap_minutes>=1559

** FT: Blacks
** Below median
ttest ft_blacks_ibrc = ft_blacks_online if timegap_minutes<1559
** at or above median
ttest ft_blacks_ibrc = ft_blacks_online if timegap_minutes>=1559

** FT: Democrats
** Below median
ttest ft_dems_ibrc=ft_dems_online if timegap_minutes<1559
** at or above median
ttest ft_dems_ibrc=ft_dems_online if timegap_minutes>=1559

** FT: Evangelicals
** Below median
ttest ft_evanelicals_ibrc = ft_evanelicals_online if timegap_minutes<1559
** at or above median
ttest ft_evanelicals_ibrc = ft_evanelicals_online if timegap_minutes>=1559

** FT: Gay men and lesbians
** Below median
ttest ft_gays_ibrc = ft_gays_online if timegap_minutes<1559
** at or above median
ttest ft_gays_ibrc = ft_gays_online if timegap_minutes>=1559

** FT: Muslims
** Below median
ttest ft_muslims_ibrc = ft_muslims_online if timegap_minutes<1559
** at or above median
ttest ft_muslims_ibrc = ft_muslims_online if timegap_minutes>=1559

** FT: Republicans
** Below median
ttest ft_reps_ibrc = ft_reps_online if timegap_minutes<1559
** at or above median
ttest ft_reps_ibrc = ft_reps_online if timegap_minutes>=1559


** Mean issues mentioned
** Below median
ttest mip_ibrc_problemcount = mip_online_problemcount if timegap_minutes<1559
** at or above median
ttest mip_ibrc_problemcount = mip_online_problemcount if timegap_minutes>=1559

** Item nonresponse
** Below median
ttest itemnonresponse_ibrc== itemnonresponse_online if timegap_minutes<1559
** at or above median
ttest itemnonresponse_ibrc== itemnonresponse_online if timegap_minutes>=1559

** Straightlining
** Below median
ttest sl_any_ibrc== sl_any_online if timegap_minutes<1559
** at or above median
ttest sl_any_ibrc== sl_any_online if timegap_minutes>=1559


log close





